Electron microscope (EM) tomography is now a proven method for the study of cellular architecture at approximately 5nm resolution. We seek to carry out work on the Golgi complex and components of the cytoskeleton and to enable high resolution EM of frozen-hydrated macromolecules. Our CORE research will pursue three technological goals. We will automate tomography, making it convenient to generate 3- D reconstructions of large volumes of cellular structure. This advance will provide both scientists at Boulder and structural biologist elsewhere with the hardware and computational tools that will allow 3-D reconstructions of extensive cellular structures, like cytomembrane systems and parts of the cytoskeleton. We will also work to improve methods for fixation by freeze-substitution and for the labeling of biological macromolecules in well preserved material. Existing methods, such as immunoEM, have important limitations, and while we will use these approaches for some work, we propose four novel approaches to labeling that may improve the resolution, ease, and sensitivity of macromolecular localizations in cells. We will also work to make the modeling of tomograms more objective and efficient, so we can extract reliable information with greater speed, producing data for a quantitative examination of reliable information with greater speed, producing data for a quantitative examination of reliable information with greater speed, producing data for a quantitative examination of variability in cellular fine structure. Each of these advances will help structural cell biology move to a new level of reliability and effectiveness in answering important scientific questions.